Is chordin a long-range- or short-range-acting factor? Roles for BMP1-related metalloproteases in chordin and BMP4 autofeedback loop regulation.

Diffusible morphogen models have been used widely to explain regional specification of tissues and body axes during animal development. The three-signal model for patterning the dorsal-ventral axis of the amphibian embryo proposes, in part, that a factor(s) secreted from Spemann's organizer is responsible for converting lateral marginal zone into more dorsal cell fates. We examine the possibility that chordin, a secreted inhibitor of bone morphogenetic protein (BMP) signaling and candidate "dorsalizing signal," is a long-range-acting factor. We show that chordin can, when overexpressed, act directly over distances of at least 450 microm in the early Xenopus embryo to create a gradient of BMP signaling. However, since lower levels of chordin can still induce secondary axes and these amounts of chordin act only locally to inhibit a BMP target gene, we suggest that chordin likely acts as a short-range signal in vivo. Furthermore, BMP1, a secreted metalloprotease that cleaves chordin protein in vitro, inhibits chordin's axis-inducing effects, suggesting that BMP1 functions to negatively regulate chordin's action in vivo. A dominant-negative mutant BMP1 blocks the in vitro cleavage of chordin protein by wild-type BMP1 and induces secondary axes when injected ventrally. We argue that BMP1 and Xolloid are probably functionally redundant metalloproteases and may have two roles in the early Xenopus embryo. One role may be to inhibit the action of low-level chordin protein expressed throughout the entire embryo and a possible second role may be to inhibit activation of a juxtacrine cell relay, thereby confining chordin's action to the organizer region preventing chordin from functioning as a long-range-acting factor. Copyright 2000 Academic Press.